Framework guide

ICE Scoring for Product Prioritization

ICE is a fast way to compare experiments when reach is unknown. Its speed is useful early, but only if the team defines what each number means before scoring.

About this guide

Prepared by LearnLive with AI assistance, then checked against the sources listed below. Published July 22, 2026.

Short answer

Score each idea from 1 to 10 for impact, confidence, and ease, then multiply the three values. Use ICE for fast experiment backlogs, not as a precise financial model.

  • Define the outcome and the 1-to-10 anchors before anyone scores ideas.
  • Ease is the inverse of effort, so easier work receives a higher number.
  • Re-score after an experiment produces evidence instead of preserving the original estimate.

How ICE scoring works

Impact asks how strongly the idea could move the chosen outcome. Confidence asks how much evidence supports that impact estimate. Ease asks how quickly and safely the team can deliver and learn. Multiplication makes a very weak factor pull the whole score down.

Relay defines 10 impact as a likely step-change in activation, 5 as a noticeable movement, and 1 as negligible. It defines 10 confidence as direct product evidence and 1 as intuition only. For ease, 10 means less than a day and 1 means a multi-team project.

ICE scores for Relay activation experiments
ExperimentImpactConfidenceEaseICE score
Weekly progress email688384
Onboarding checklist876336
AI setup assistant93381

Read the ranking as a learning order

The weekly email ranks first because it combines reasonable expected impact with strong evidence and low delivery cost. The AI setup assistant has the largest imagined impact but little evidence and much more work. ICE suggests learning from the email first, not abandoning the larger idea forever.

If two people produce very different scores, do not average them immediately. Ask which assumption differs. A disagreement about confidence may reveal missing data; a disagreement about ease may reveal hidden technical work.

Use ICE for speed, not precision

ICE fits growth and discovery backlogs where the team needs to choose the next few tests. Use RICE when the number of affected users matters and can be estimated. Use weighted scoring when the decision must balance several explicit business criteria.

Common mistakes

  • Scoring before naming the outcome the team wants to move.
  • Letting every scorer invent a different meaning for 7 or 8.
  • Using effort instead of ease, which reverses the intended direction.
  • Comparing scores created by different teams with different anchors.

Check your understanding

An idea scores impact 8, confidence 5, and ease 4. What is its ICE score?

160, because 8 x 5 x 4 = 160.

Sources

Build a course around your product decisions

Start with product prioritization, then tell LearnLive about your role, experience, and goal.

Start this topic